Alexey Dobriyan [Wed, 17 May 2017 16:31:39 +0000 (19:31 +0300)]
net: make struct dst_entry::dev first member
struct dst_entry::dev is used most often. Move it so it can be
accessed without imm8 offset on x86_64.
add/remove: 0/0 grow/shrink: 9/239 up/down: 52/-413 (-361)
function old new delta
dst_rcu_free 126 138 +12
fnhe_flush_routes 211 219 +8
rt_set_nexthop 747 754 +7
rt_cache_route 85 91 +6
rt6_release 209 215 +6
dst_release 107 111 +4
dst_destroy_rcu 29 33 +4
dn_dst_check_expire 329 333 +4
dn_insert_route 484 485 +1
xfrm_resolve_and_create_bundle 2991 2990 -1
...
ip_route_me_harder 1163 1157 -6
__ip_append_data.isra 2730 2724 -6
ip6_forward 3052 3045 -7
callforward_do_filter 659 651 -8
dst_gc_task 571 549 -22
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 18 May 2017 14:28:49 +0000 (10:28 -0400)]
Merge branch 'fsl_ucc_hdlc-enhancements'
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:39 +0000 (17:24 +0200)]
powerpc/85xx/kmcent2: use hdlc busmode for UCC1
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:38 +0000 (17:24 +0200)]
net/wan/fsl_ucc_hdlc: add hdlc-bus support
This adds support for hdlc-bus mode to the fsl_ucc_hdlc driver. This can
be enabled with the "fsl,hdlc-bus" property in the DTS node of the
corresponding ucc.
This aligns the configuration of the UPSMR and GUMR registers to what is
done in our ucc_hdlc driver (that only support hdlc-bus mode) and with
the QuickEngine's documentation for hdlc-bus mode.
GUMR/SYNL is set to AUTO for the busmode as in this case the CD signal
is ignored. The brkpt_support is enabled to set the HBM1 bit in the
CMXUCR register to configure an open-drain connected HDLC bus.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Zhao Qiang <qiang.zhao@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:37 +0000 (17:24 +0200)]
fsl/qe: add bit description for SYNL register for GUMR
Add the bitmask for the two bit SYNL register according to the QUICK
Engine Reference Manual.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Zhao Qiang <qiang.zhao@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:36 +0000 (17:24 +0200)]
net/wan/fsl_ucc_hdlc: call qe_setbrg only for loopback mode
We can't assume that we are always in loopback mode if rx and tx clock
have the same clock source. If we want to use HDLC busmode we also have
the same clock source but we are not in loopback mode. So move the
setting of the baudrate generator after the check for property for the
loopback mode.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Zhao Qiang <qiang.zhao@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:35 +0000 (17:24 +0200)]
net/wan/fsl_ucc_hdlc: fix incorrect memory allocation
We need space for the struct qe_bd and not for a pointer to this struct.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Zhao Qiang <qiang.zhao@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:34 +0000 (17:24 +0200)]
net/wan/fsl_ucc_hdlc: fix wrong indentation
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Zhao Qiang <qiang.zhao@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:33 +0000 (17:24 +0200)]
net/wan/fsl_ucc_hdlc: fix unitialized variable warnings
This fixes the following compiler warnings:
drivers/net/wan/fsl_ucc_hdlc.c: In function 'ucc_hdlc_poll':
warning: 'skb' may be used uninitialized in this function
[-Wmaybe-uninitialized]
skb->mac_header = skb->data - skb->head;
and
drivers/net/wan/fsl_ucc_hdlc.c: In function 'ucc_hdlc_probe':
drivers/net/wan/fsl_ucc_hdlc.c:1127:3: warning: 'utdm' may be used
uninitialized in this function [-Wmaybe-uninitialized]
kfree(utdm);
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Zhao Qiang <qiang.zhao@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Brunck [Wed, 17 May 2017 15:24:32 +0000 (17:24 +0200)]
net/wan/fsl_ucc_hdlc: cleanup debug traces
Some of the tracing seems to be remaining traces for basic driver
development. They can be removed now, as they cause noisy printouts.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
Cc: Zhao Qiang <qiang.zhao@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alexey Dobriyan [Wed, 17 May 2017 10:30:44 +0000 (13:30 +0300)]
net: make struct net_device::tx_queue_len unsigned int
4 billion packet queue is something unthinkable so use 32-bit value
for now.
Space savings on x86_64:
add/remove: 0/0 grow/shrink: 3/70 up/down: 16/-131 (-115)
function old new delta
change_tx_queue_len 94 108 +14
qdisc_create 1176 1177 +1
alloc_netdev_mqs 1124 1125 +1
xenvif_alloc 533 532 -1
x25_asy_setup 167 166 -1
...
tun_queue_resize 945 940 -5
pfifo_fast_enqueue 167 162 -5
qfq_init_qdisc 168 158 -10
tap_queue_resize 810 799 -11
transmit 719 698 -21
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Colin Ian King [Wed, 17 May 2017 08:50:36 +0000 (09:50 +0100)]
udp: make function udp_skb_dtor_locked static
Function udp_skb_dtor_locked does not need to be in global scope
so make it static to fix sparse warning:
net/ipv4/udp.c: warning: symbol 'udp_skb_dtor_locked' was not
declared. Should it be static?
Fixes:
6dfb4367cd911d ("udp: keep the sk_receive_queue held when splicing")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 18 May 2017 14:07:42 +0000 (10:07 -0400)]
Merge branch 'vhost_net-rx-batch-dequeuing'
Jason Wang says:
====================
vhost_net rx batch dequeuing
This series tries to implement rx batching for vhost-net. This is done
by batching the dequeuing from skb_array which was exported by
underlayer socket and pass the sbk back through msg_control to finish
userspace copying. This is also the requirement for more batching
implemention on rx path.
Tests shows at most 7.56% improvment bon rx pps on top of batch
zeroing and no obvious changes for TCP_STREAM/TCP_RR result.
Please review.
Thanks
Changes from V4:
- drop batch zeroing patch
- renew the performance numbers
- move skb pointer array out of vhost_net structure
Changes from V3:
- add batch zeroing patch to fix the build warnings
Changes from V2:
- rebase to net-next HEAD
- use unconsume helpers to put skb back on releasing
- introduce and use vhost_net internal buffer helpers
- renew performance numbers on top of batch zeroing
Changes from V1:
- switch to use for() in __ptr_ring_consume_batched()
- rename peek_head_len_batched() to fetch_skbs()
- use skb_array_consume_batched() instead of
skb_array_consume_batched_bh() since no consumer run in bh
- drop the lockless peeking patch since skb_array could be resized, so
it's not safe to call lockless one
====================
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:45 +0000 (12:14 +0800)]
vhost_net: try batch dequing from skb array
We used to dequeue one skb during recvmsg() from skb_array, this could
be inefficient because of the bad cache utilization and spinlock
touching for each packet. This patch tries to batch them by calling
batch dequeuing helpers explicitly on the exported skb array and pass
the skb back through msg_control for underlayer socket to finish the
userspace copying. Batch dequeuing is also the requirement for more
batching improvement on receive path.
Tests were done by pktgen on tap with XDP1 in guest. Host is Intel(R)
Xeon(R) CPU E5-2650 0 @ 2.00GHz.
rx batch | pps
0 2.25Mpps
1 2.33Mpps (+3.56%)
4 2.33Mpps (+3.56%)
16 2.35Mpps (+4.44%)
64 2.42Mpps (+7.56%) <- Default rx batching
128 2.40Mpps (+6.67%)
256 2.38Mpps (+5.78%)
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:44 +0000 (12:14 +0800)]
tap: support receiving skb from msg_control
This patch makes tap_recvmsg() can receive from skb from its caller
through msg_control. Vhost_net will be the first user.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:43 +0000 (12:14 +0800)]
tun: support receiving skb through msg_control
This patch makes tun_recvmsg() can receive from skb from its caller
through msg_control. Vhost_net will be the first user.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:42 +0000 (12:14 +0800)]
tap: export skb_array
This patch exports skb_array through tap_get_skb_array(). Caller can
then manipulate skb array directly.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:41 +0000 (12:14 +0800)]
tun: export skb_array
This patch exports skb_array through tun_get_skb_array(). Caller can
then manipulate skb array directly.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:40 +0000 (12:14 +0800)]
skb_array: introduce batch dequeuing
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:39 +0000 (12:14 +0800)]
ptr_ring: introduce batch dequeuing
This patch introduce a batched version of consuming, consumer can
dequeue more than one pointers from the ring at a time. We don't care
about the reorder of reading here so no need for compiler barrier.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Wed, 17 May 2017 04:14:38 +0000 (12:14 +0800)]
skb_array: introduce skb_array_unconsume
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Michael S. Tsirkin [Wed, 17 May 2017 04:14:37 +0000 (12:14 +0800)]
ptr_ring: add ptr_ring_unconsume
Applications that consume a batch of entries in one go
can benefit from ability to return some of them back
into the ring.
Add an API for that - assuming there's space. If there's no space
naturally can't do this and have to drop entries, but this implies ring
is full so we'd likely drop some anyway.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 17 May 2017 20:27:52 +0000 (16:27 -0400)]
Merge branch 'phy-marvell-cleanups'
Andrew Lunn says:
====================
net: phy: marvell: Checkpatch cleanup
I will be contributing a few new features to the Marvell PHY driver
soon. Start by making the code mostly checkpatch clean. There should
not be any functional changes. Just comments set into the correct
format, missing blank lines, turn some comparisons around, and
refactoring to reduce indentation depth.
There is still one camel in the code, but it actually makes sense, so
leave it in piece.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Wed, 17 May 2017 01:26:04 +0000 (03:26 +0200)]
net: phy: marvell: checkpatch - Fix remaining long lines
Fold lines longer than 80 characters
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Wed, 17 May 2017 01:26:03 +0000 (03:26 +0200)]
net: phy: marvell: Add helpers to get/set page
Makes the code a bit more readable, and solves quite a few checkpatch
warnings of lines longer than 80 characters.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Wed, 17 May 2017 01:26:02 +0000 (03:26 +0200)]
net: phy: marvell: Refactor some bigger functions
Break big functions up by using a number of smaller helper
function. Solves some of the over 80 lines warnings, by reducing the
indentation level.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Wed, 17 May 2017 01:26:01 +0000 (03:26 +0200)]
net: phy: marvell: Checkpatch - assignments and comparisons
Avoid multiple assignments
Comparisons should place the constant on the right side of the test
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Wed, 17 May 2017 01:26:00 +0000 (03:26 +0200)]
net: phy: marvell: Checkpatch - Missing or extra blank lines
Remove the extra blank lines, add one in where recommended.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Wed, 17 May 2017 01:25:59 +0000 (03:25 +0200)]
net: phy: Marvell: checkpatch - Comments
Use net style comment blocks, and wrap one block with long lines.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 17 May 2017 20:06:03 +0000 (16:06 -0400)]
Merge branch 'tcp-TCP-TS-option-use-1-ms-clock'
Eric Dumazet says:
====================
tcp: TCP TS option use 1 ms clock
TCP Timestamps option is defined in RFC 7323
Traditionally on linux, it has been tied to the internal
'jiffy' variable, because it had been a cheap and good enough
generator.
Unfortunately some distros use HZ=250 or even HZ=100 leading
to not very useful TCP timestamps.
For TCP flows in the DC, Google has used usec resolution for more
than two years with great success [1].
RCVBUF autotuning is more precise.
This series converts tp->tcp_mstamp to a plain u64 value storing
a 1 usec TCP clock.
This choice will allow us to upstream the 1 usec TS option as
discussed in IETF 97.
Kathleen Nichols [2] and others advocate for 1ms TS clocks for
network analysis. (1ms being the lowest value supported by RFC 7323.)
[1] https://www.ietf.org/proceedings/97/slides/slides-97-tcpm-tcp-options-for-low-latency-00.pdf
[2] http://netseminar.stanford.edu/seminars/02_02_17.pdf
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:14 +0000 (14:00 -0700)]
tcp: switch TCP TS option (RFC 7323) to 1ms clock
TCP Timestamps option is defined in RFC 7323
Traditionally on linux, it has been tied to the internal
'jiffies' variable, because it had been a cheap and good enough
generator.
For TCP flows on the Internet, 1 ms resolution would be much better
than 4ms or 10ms (HZ=250 or HZ=100 respectively)
For TCP flows in the DC, Google has used usec resolution for more
than two years with great success [1]
Receive size autotuning (DRS) is indeed more precise and converges
faster to optimal window size.
This patch converts tp->tcp_mstamp to a plain u64 value storing
a 1 usec TCP clock.
This choice will allow us to upstream the 1 usec TS option as
discussed in IETF 97.
[1] https://www.ietf.org/proceedings/97/slides/slides-97-tcpm-tcp-options-for-low-latency-00.pdf
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:13 +0000 (14:00 -0700)]
tcp: replace misc tcp_time_stamp to tcp_jiffies32
After this patch, all uses of tcp_time_stamp will require
a change when we introduce 1 ms and/or 1 us TCP TS option.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:12 +0000 (14:00 -0700)]
tcp_lp: cache tcp_time_stamp
tcp_time_stamp will become slightly more expensive soon,
cache its value.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:11 +0000 (14:00 -0700)]
tcp_westwood: use tcp_jiffies32 instead of tcp_time_stamp
This CC does not need 1 ms tcp_time_stamp and can use
the jiffy based 'timestamp'.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:10 +0000 (14:00 -0700)]
tcp: use tcp_jiffies32 in __tcp_oow_rate_limited()
This place wants to use tcp_jiffies32, this is good enough.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:09 +0000 (14:00 -0700)]
tcp: uses jiffies_32 to feed tp->chrono_start
tcp_time_stamp will no longer be tied to jiffies.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:08 +0000 (14:00 -0700)]
tcp: use tcp_jiffies32 to feed probe_timestamp
Use tcp_jiffies32 instead of tcp_time_stamp, since
tcp_time_stamp will soon be only used for TCP TS option.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:07 +0000 (14:00 -0700)]
tcp: use tcp_jiffies32 for rcv_tstamp and lrcvtime
Use tcp_jiffies32 instead of tcp_time_stamp, since
tcp_time_stamp will soon be only used for TCP TS option.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:06 +0000 (14:00 -0700)]
tcp: bic, cubic: use tcp_jiffies32 instead of tcp_time_stamp
Use tcp_jiffies32 instead of tcp_time_stamp, since
tcp_time_stamp will soon be only used for TCP TS option.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:05 +0000 (14:00 -0700)]
tcp_bbr: use tcp_jiffies32 instead of tcp_time_stamp
Use tcp_jiffies32 instead of tcp_time_stamp, since
tcp_time_stamp will soon be only used for TCP TS option.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:04 +0000 (14:00 -0700)]
tcp: use tcp_jiffies32 to feed tp->snd_cwnd_stamp
Use tcp_jiffies32 instead of tcp_time_stamp to feed
tp->snd_cwnd_stamp.
tcp_time_stamp will soon be a litle bit more expensive
than simply reading 'jiffies'.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:03 +0000 (14:00 -0700)]
tcp: use tcp_jiffies32 to feed tp->lsndtime
Use tcp_jiffies32 instead of tcp_time_stamp to feed
tp->lsndtime.
tcp_time_stamp will soon be a litle bit more expensive
than simply reading 'jiffies'.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:02 +0000 (14:00 -0700)]
dccp: do not use tcp_time_stamp
Use our own macro instead of abusing tcp_time_stamp
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:01 +0000 (14:00 -0700)]
tcp: introduce tcp_jiffies32
We abuse tcp_time_stamp for two different cases :
1) base to generate TCP Timestamp options (RFC 7323)
2) A 32bit version of jiffies since some TCP fields
are 32bit wide to save memory.
Since we want in the future to have 1ms TCP TS clock,
regardless of HZ value, we want to cleanup things.
tcp_jiffies32 is the truncated jiffies value,
which will be used only in places where we want a 'host'
timestamp.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 21:00:00 +0000 (14:00 -0700)]
tcp: use tp->tcp_mstamp in output path
Idea is to later convert tp->tcp_mstamp to a full u64 counter
using usec resolution, so that we can later have fine
grained TCP TS clock (RFC 7323), regardless of HZ value.
We try to refresh tp->tcp_mstamp only when necessary.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 17 May 2017 20:03:16 +0000 (16:03 -0400)]
sch_dsmark: Fix uninitialized variable warning.
We still need to initialize err to -EINVAL for
the case where 'opt' is NULL in dsmark_init().
Fixes:
6529eaba33f0 ("net: sched: introduce tcf block infractructure")
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 17 May 2017 19:22:15 +0000 (15:22 -0400)]
Merge branch 'net-sched-multichain-filters'
Jiri Pirko says:
====================
net: sched: introduce multichain support for filters
Currently, each classful qdisc holds one chain of filters.
This chain is traversed and each filter could be matched on, which
may lead to execution of list of actions. One of such action
could be "reclassify", which would "reset" the processing of the
filter chain.
So this filter chain could be looked at as a flat table.
Sometimes it is convenient for user to configure a hierarchy
of tables. Example usecase is encapsulation.
Hierarchy of tables is a common way how it is done in HW pipelines.
So it is much more convenient to offload this.
This patchset contains two major patches:
8/10 - This patch introduces the support for having multiple
chains of filters.
10/10 - This patch adds new control action to allow going to specified chain
The rest of the patches are smaller or bigger depencies of those 2.
Please see individual patch descriptions for details.
Corresponding iproute2 patches are appended as a reply to this cover letter.
Simple example:
$ tc qdisc add dev eth0 ingress
$ tc filter add dev eth0 parent ffff: protocol ip pref 33 flower dst_mac 52:54:00:3d:c7:6d action goto chain 11
$ tc filter add dev eth0 parent ffff: protocol ip pref 22 chain 11 flower dst_ip 192.168.40.1 action drop
$ tc filter show dev eth0 root
filter parent ffff: protocol ip pref 33 flower chain 0
filter parent ffff: protocol ip pref 33 flower chain 0 handle 0x1
dst_mac 52:54:00:3d:c7:6d
eth_type ipv4
action order 1: gact action goto chain 11
random type none pass val 0
index 2 ref 1 bind 1
filter parent ffff: protocol ip pref 22 flower chain 11
filter parent ffff: protocol ip pref 22 flower chain 11 handle 0x1
eth_type ipv4
dst_ip 192.168.40.1
action order 1: gact action drop
random type none pass val 0
index 3 ref 1 bind 1
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:08:03 +0000 (11:08 +0200)]
net: sched: add termination action to allow goto chain
Introduce new type of termination action called "goto_chain". This allows
user to specify a chain to be processed. This action type is
then processed as a return value in tcf_classify loop in similar
way as "reclassify" is, only it does not reset to the first filter
in chain but rather reset to the first filter of the desired chain.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:08:02 +0000 (11:08 +0200)]
net: sched: push tp down to action init
Tp pointer will be needed by the next patch in order to get the chain.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:08:01 +0000 (11:08 +0200)]
net: sched: introduce multichain support for filters
Instead of having only one filter per block, introduce a list of chains
for every block. Create chain 0 by default. UAPI is extended so the user
can specify which chain he wants to change. If the new attribute is not
specified, chain 0 is used. That allows to maintain backward
compatibility. If chain does not exist and user wants to manipulate with
it, new chain is created with specified index. Also, when last filter is
removed from the chain, the chain is destroyed.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:08:00 +0000 (11:08 +0200)]
net: sched: push chain dump to a separate function
Since there will be multiple chains to dump, push chain dumping code to
a separate function.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:07:59 +0000 (11:07 +0200)]
net: sched: introduce helpers to work with filter chains
Introduce struct tcf_chain object and set of helpers around it. Wraps up
insertion, deletion and search in the filter chain.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:07:58 +0000 (11:07 +0200)]
net: sched: move TC_H_MAJ macro call into tcf_auto_prio
Call the helper from the function rather than to always adjust the
return value of the function.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:07:57 +0000 (11:07 +0200)]
net: sched: replace nprio by a bool to make the function more readable
The use of "nprio" variable in tc_ctl_tfilter is a bit cryptic and makes
a reader wonder what is going on for a while. So help him to understand
this priority allocation dance a litte bit better.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:07:56 +0000 (11:07 +0200)]
net: sched: rename tcf_destroy_chain helper
Make the name consistent with the rest of the helpers around.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:07:55 +0000 (11:07 +0200)]
net: sched: introduce tcf block infractructure
Currently, the filter chains are direcly put into the private structures
of qdiscs. In order to be able to have multiple chains per qdisc and to
allow filter chains sharing among qdiscs, there is a need for common
object that would hold the chains. This introduces such object and calls
it "tcf_block".
Helpers to get and put the blocks are provided to be called from
individual qdisc code. Also, the original filter_list pointers are left
in qdisc privs to allow the entry into tcf_block processing without any
added overhead of possible multiple pointer dereference on fast path.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 17 May 2017 09:07:54 +0000 (11:07 +0200)]
net: sched: move tc_classify function to cls_api.c
Move tc_classify function to cls_api.c where it belongs, rename it to
fit the namespace.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 17 May 2017 19:19:40 +0000 (15:19 -0400)]
Merge branch 'dsa-sort'
Andrew Lunn says:
====================
net: dsa: Sort various lists
As we gain more DSA drivers and tagging protocols, the lists are
getting a bit unruly. Do some sorting.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Tue, 16 May 2017 20:40:08 +0000 (22:40 +0200)]
drivers: net: DSA: Sort drivers
With more drivers being added, it is time to sort the drivers to
impose some order.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Tue, 16 May 2017 20:40:07 +0000 (22:40 +0200)]
net: dsa: Sort DSA tagging protocol drivers
With more tag protocols being added, regain some order by sorting the
entries in various places.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Felix Manlunas [Tue, 16 May 2017 18:28:00 +0000 (11:28 -0700)]
liquidio: fix PF falsely indicating success at setting MAC address of a nonexistent VF
In the function assigned to .ndo_set_vf_mac, check the validity of the
vfidx argument before proceeding to tell the firmware to set the VF MAC
address.
Signed-off-by: Felix Manlunas <felix.manlunas@cavium.com>
Signed-off-by: Derek Chickles <derek.chickles@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rick Farrington [Tue, 16 May 2017 18:14:50 +0000 (11:14 -0700)]
liquidio: fix insmod failure when multiple NICs are plugged in
When multiple liquidio NICs are plugged in, the first insmod of the PF
driver succeeds. But after an rmmod, a subsequent insmod fails. Reason is
during rmmod, the PF driver resets the Octeon of only one of the NICs; it
neglects to reset the Octeons of the other NICs.
Fix the insmod failure by adding the missing Octeon resets at rmmod. Keep
a per-NIC refcount that indicates the number of active PFs in a given NIC.
When the refcount goes to zero, then reset the Octeon of that NIC.
Signed-off-by: Rick Farrington <ricardo.farrington@cavium.com>
Signed-off-by: Felix Manlunas <felix.manlunas@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vivien Didelot [Tue, 16 May 2017 18:10:33 +0000 (14:10 -0400)]
net: dsa: store CPU port pointer in the tree
A dsa_switch_tree instance holds a dsa_switch pointer and a port index
to identify the switch port to which the CPU is attached.
Now that the DSA layer has a dsa_port structure to hold this data, use
it to point the switch CPU port.
This patch simply substitutes s/dst->cpu_switch/dst->cpu_dp->ds/ and
s/dst->cpu_port/dst->cpu_dp->index/.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 17 May 2017 18:06:56 +0000 (14:06 -0400)]
Merge branch 'mlxsw-Preparations-for-restructuring'
Jiri Pirko says:
====================
mlxsw: Preparations for restructuring
This patchset doesn't introduce any functional changes and merely meant
to make the code base more receptive for upcoming restructuring.
The first six patches mainly shuffle code in order to reduce the scope of
structs that shouldn't be defined in the main driver header. Most of them
will be later expanded, so it makes sense to correctly place them now.
The last patches mostly simplify bridge-related functions, so that they
could be more easily modified later on.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:35 +0000 (19:38 +0200)]
mlxsw: spectrum: Default ports to non-virtual mode
In virtual mode, packets are classified to FIDs based on their ingress
port and VLAN whereas in non-virtual mode only the VLAN is taken into
account.
Currently ports are initialized to use virtual mode due to the presence
of the PVID vPort. However, we're going to transition ports between both
modes based on the FIDs they use and not merely based on the presence on
a VLAN upper. Therefore, during initialization, no mode will be
explicitly set.
Since the Programmer's Reference Manual (PRM) doesn't specify a default,
explicitly set the port to non-virtual mode and later transition the
port between both modes based on the FIDs it uses.
In a follow-up patchset, this step will be moved to the common FID core
where it logically belongs.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:34 +0000 (19:38 +0200)]
mlxsw: spectrum: Move PVID code to appropriate place
PVID is a port attribute and should therefore reside in the main driver
file and not the switchdev specific one.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:33 +0000 (19:38 +0200)]
mlxsw: spectrum_switchdev: Don't batch learning operations
We no longer batch VLAN operations, so there's no need to set the
learning state for a range of VLANs.
Use a common function to set the learning state for a Port-VLAN, thereby
making the code saner more receptive for upcoming changes.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:32 +0000 (19:38 +0200)]
mlxsw: spectrum_switchdev: Don't batch STP operations
Simplify the code by using the common function that sets an STP state
for a Port-VLAN and remove the existing one that tries to batch it for
several VLANs.
This will help us in a follow-up patchset to introduce a unified
infrastructure for bridge ports, regardless if the bridge is VLAN-aware
or not.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:31 +0000 (19:38 +0200)]
mlxsw: spectrum_switchdev: Don't batch VLAN operations
switchdev's VLAN object has the ability to describe a range of VLAN IDs,
but this is only used when VLAN operations are done using the SELF flag,
which is something we would like to remove as it allows one to bypass
the bridge driver.
Do VLAN operations on a per-VLAN basis, thereby simplifying the code and
preparing it for refactoring in a follow-up patchset.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:30 +0000 (19:38 +0200)]
mlxsw: spectrum_switchdev: Remove redundant check
Since commit
97c242902c20 ("switchdev: Execute bridge ndos only for
bridge ports") switchdev code checks that port is bridged, so no need to
perform the same check in the driver.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:29 +0000 (19:38 +0200)]
mlxsw: spectrum_router: Initialize RIFs in a separate function
The router interfaces (RIFs) array is currently initialized together
with the general router configuration. However, in a follow-up patchset
we're going to introduce a common RIF core that will require us to
initialize more RIF constructs, so move the RIF initialization to its
own function.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:28 +0000 (19:38 +0200)]
mlxsw: spectrum_router: Move FIB notification block to router struct
The FIB notification block logically belongs inside the router specific
struct, so move it there.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:27 +0000 (19:38 +0200)]
mlxsw: spectrum_router: Move RIFs array to its rightful place
The router interfaces (RIFs) array is of no interest to code outside the
routing realm, so declare it inside the router specific struct instead
of the chip-wide one.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:26 +0000 (19:38 +0200)]
mlxsw: spectrum_switchdev: Reduce scope of bridge struct
Some attributes in the global chip struct are only relevant for bridge
operation, so encapsulate them in their own struct that isn't exposed to
non-bridge code.
This will also help us later, when we add more bridge-specific
attributes.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:25 +0000 (19:38 +0200)]
mlxsw: spectrum_router: Reduce scope of router struct
In a similar fashion to previous patch, the router structure
('mlxsw_sp_router') doesn't need to be accessible to anyone, but the
router code located at spectrum_router.c
Make this apparent and reduce its scope by defining it there.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Tue, 16 May 2017 17:38:24 +0000 (19:38 +0200)]
mlxsw: spectrum_buffer: Reduce scope of shared buffer struct
The shared buffer structure ('mlxsw_sp_sb') doesn't need to be
accessible to anyone, but the shared buffer code located at
spectrum_buffers.c
Make this apparent and reduce its scope by defining it there.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ganesh Goudar [Tue, 16 May 2017 16:09:05 +0000 (21:39 +0530)]
cxgb4: add new T5 pci device id
Signed-off-by: Ganesh Goudar <ganeshgr@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ganesh Goudar [Tue, 16 May 2017 15:47:42 +0000 (21:17 +0530)]
cxgb4: reduce resource allocation in kdump kernel
When is_kdump_kernel() is true, reduce memory footprint of
cxgb4 by using a single "Queue Set".
Signed-off-by: Ganesh Goudar <ganeshgr@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Christoph Hellwig [Tue, 16 May 2017 14:21:46 +0000 (16:21 +0200)]
liquidio: use pcie_flr instead of duplicating it
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Felix Manlunas <felix.manlunas@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Tue, 16 May 2017 16:29:11 +0000 (18:29 +0200)]
net: phy: Remove residual magic from PHY drivers
commit
fa8cddaf903c ("net phylib: Remove unnecessary condition check in phy")
removed the only place where the PHY flag PHY_HAS_MAGICANEG was
checked. But it left the flag being set in the drivers. Remove the flag.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Leon Romanovsky [Tue, 16 May 2017 12:20:56 +0000 (15:20 +0300)]
bnx2x: Remove open coded carrier check
There is inline function to test if carrier present,
so it makes open-coded solution redundant.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Acked-by: Yuval Mintz <Yuval.Mintz@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 16 May 2017 11:24:36 +0000 (04:24 -0700)]
tcp: internal implementation for pacing
BBR congestion control depends on pacing, and pacing is
currently handled by sch_fq packet scheduler for performance reasons,
and also because implemening pacing with FQ was convenient to truly
avoid bursts.
However there are many cases where this packet scheduler constraint
is not practical.
- Many linux hosts are not focusing on handling thousands of TCP
flows in the most efficient way.
- Some routers use fq_codel or other AQM, but still would like
to use BBR for the few TCP flows they initiate/terminate.
This patch implements an automatic fallback to internal pacing.
Pacing is requested either by BBR or use of SO_MAX_PACING_RATE option.
If sch_fq happens to be in the egress path, pacing is delegated to
the qdisc, otherwise pacing is done by TCP itself.
One advantage of pacing from TCP stack is to get more precise rtt
estimations, and less work done from TX completion, since TCP Small
queue limits are not generally hit. Setups with single TX queue but
many cpus might even benefit from this.
Note that unlike sch_fq, we do not take into account header sizes.
Taking care of these headers would add additional complexity for
no practical differences in behavior.
Some performance numbers using 800 TCP_STREAM flows rate limited to
~48 Mbit per second on 40Gbit NIC.
If MQ+pfifo_fast is used on the NIC :
$ sar -n DEV 1 5 | grep eth
14:48:44 eth0 725743.00
2932134.00 46776.76
4335184.68 0.00 0.00 1.00
14:48:45 eth0 725349.00
2932112.00 46751.86
4335158.90 0.00 0.00 0.00
14:48:46 eth0 725101.00
2931153.00 46735.07
4333748.63 0.00 0.00 0.00
14:48:47 eth0 725099.00
2931161.00 46735.11
4333760.44 0.00 0.00 1.00
14:48:48 eth0 725160.00
2931731.00 46738.88
4334606.07 0.00 0.00 0.00
Average: eth0 725290.40
2931658.20 46747.54
4334491.74 0.00 0.00 0.40
$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
4 0 0
259825920 45644
2708324 0 0 21 2 247 98 0 0 100 0 0
4 0 0
259823744 45644
2708356 0 0 0 0
2400825 159843 0 19 81 0 0
0 0 0
259824208 45644
2708072 0 0 0 0
2407351 159929 0 19 81 0 0
1 0 0
259824592 45644
2708128 0 0 0 0
2405183 160386 0 19 80 0 0
1 0 0
259824272 45644
2707868 0 0 0 32
2396361 158037 0 19 81 0 0
Now use MQ+FQ :
lpaa23:~# echo fq >/proc/sys/net/core/default_qdisc
lpaa23:~# tc qdisc replace dev eth0 root mq
$ sar -n DEV 1 5 | grep eth
14:49:57 eth0 678614.00
2727930.00 43739.13
4033279.14 0.00 0.00 0.00
14:49:58 eth0 677620.00
2723971.00 43674.69
4027429.62 0.00 0.00 1.00
14:49:59 eth0 676396.00
2719050.00 43596.83
4020125.02 0.00 0.00 0.00
14:50:00 eth0 675197.00
2714173.00 43518.62
4012938.90 0.00 0.00 1.00
14:50:01 eth0 676388.00
2719063.00 43595.47
4020171.64 0.00 0.00 0.00
Average: eth0 676843.00
2720837.40 43624.95
4022788.86 0.00 0.00 0.40
$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
2 0 0
259832240 46008
2710912 0 0 21 2 223 192 0 1 99 0 0
1 0 0
259832896 46008
2710744 0 0 0 0
1702206 198078 0 17 82 0 0
0 0 0
259830272 46008
2710596 0 0 0 0
1696340 197756 1 17 83 0 0
4 0 0
259829168 46024
2710584 0 0 16 0
1688472 197158 1 17 82 0 0
3 0 0
259830224 46024
2710408 0 0 0 0
1692450 197212 0 18 82 0 0
As expected, number of interrupts per second is very different.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Van Jacobson <vanj@google.com>
Cc: Jerry Chu <hkchu@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 16 May 2017 19:41:31 +0000 (15:41 -0400)]
Merge branch 'udp-scalability-improvements'
Paolo Abeni says:
====================
udp: scalability improvements
This patch series implement an idea suggested by Eric Dumazet to
reduce the contention of the udp sk_receive_queue lock when the socket is
under flood.
An ancillary queue is added to the udp socket, and the socket always
tries first to read packets from such queue. If it's empty, we splice
the content from sk_receive_queue into the ancillary queue.
The first patch introduces some helpers to keep the udp code small, and the
following two implement the ancillary queue strategy. The code is split
to hopefully help the reviewing process.
The measured overall gain under udp flood is up to the 30% depending on
the numa layout and the number of ingress queue used by the relevant nic.
The performance numbers have been gathered using pktgen as sender, with 64
bytes packets, random src port on a host b2b connected via a 10Gbs link
with the dut.
The receiver used the udp_sink program by Jesper [1] and an h/w l4 rx hash on
the ingress nic, so that the number of ingress nic rx queues hit by the udp
traffic could be controlled via ethtool -L.
The udp_sink program was bound to the first idle cpu, to get more
stable numbers.
On a single numa node receiver:
nic rx queues vanilla patched kernel
1 1820 kpps 1900 kpps
2 1950 kpps 2500 kpps
16 1670 kpps 2120 kpps
When using a single nic rx queue, busy polling was also enabled,
elsewhere, in the above scenario, the bh processing becomes the bottle-neck
and this produces large artifacts in the measured performances (e.g.
improving the udp sink run time, decreases the overall tput, since more
action from the scheduler comes into play).
[1] https://github.com/netoptimizer/network-testing/blob/master/src/udp_sink.c
v1 -> v2:
Patches 1/3 and 2/3 are unchanged, in patch 3/3 the rx_queue_lock_held param
of udp_rmem_release() is now a bool.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Tue, 16 May 2017 09:20:15 +0000 (11:20 +0200)]
udp: keep the sk_receive_queue held when splicing
On packet reception, when we are forced to splice the
sk_receive_queue, we can keep the related lock held, so
that we can avoid re-acquiring it, if fwd memory
scheduling is required.
v1 -> v2:
the rx_queue_lock_held param in udp_rmem_release() is
now a bool
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Tue, 16 May 2017 09:20:14 +0000 (11:20 +0200)]
udp: use a separate rx queue for packet reception
under udp flood the sk_receive_queue spinlock is heavily contended.
This patch try to reduce the contention on such lock adding a
second receive queue to the udp sockets; recvmsg() looks first
in such queue and, only if empty, tries to fetch the data from
sk_receive_queue. The latter is spliced into the newly added
queue every time the receive path has to acquire the
sk_receive_queue lock.
The accounting of forward allocated memory is still protected with
the sk_receive_queue lock, so udp_rmem_release() needs to acquire
both locks when the forward deficit is flushed.
On specific scenarios we can end up acquiring and releasing the
sk_receive_queue lock multiple times; that will be covered by
the next patch
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Tue, 16 May 2017 09:20:13 +0000 (11:20 +0200)]
net/sock: factor out dequeue/peek with offset code
And update __sk_queue_drop_skb() to work on the specified queue.
This will help the udp protocol to use an additional private
rx queue in a later patch.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 16 May 2017 16:59:05 +0000 (12:59 -0400)]
Merge branch 'nfp-LSO-checksum-and-XDP-datapath-updates'
Jakub Kicinski says:
====================
nfp: LSO, checksum and XDP datapath updates
This series introduces a number of refinements to standard features
like LSO and checksum offload. Three major features are support for
CHECKSUM_COMPLETE, refinement of TSO handling and another small speed
up for XDP TX. This series also switches from depending on some
app FW<>driver ABI versions to heavier use of capabilities.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski [Tue, 16 May 2017 00:55:23 +0000 (17:55 -0700)]
nfp: eliminate an if statement in calculation of completed frames
Given that our rings are always a power of 2, we can simplify the
calculation of number of completed TX descriptors by using masking
instead of if statement based on whether the index have wrapped
or not.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski [Tue, 16 May 2017 00:55:22 +0000 (17:55 -0700)]
nfp: add a helper for wrapping descriptor index
We have a number of places where we calculate the descriptor
index based on a value which may have overflown. Create a
macro for masking with the ring size.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski [Tue, 16 May 2017 00:55:21 +0000 (17:55 -0700)]
nfp: complete the XDP TX ring only when it's full
Since XDP TX ring holds "spare" RX buffers anyway, we don't have to
rush the completion. We can wait until ring fills up completely
before trying to reclaim buffers. If RX poll has ended an no
buffer has been queued for XDP TX we have no guarantee we will see
another interrupt, so run the reclaim there as well, to make sure
TX statistics won't become stale.
This should help us reclaim more buffers per single queue controller
register read.
Note that the XDP completion is very trivial, it only adds up
the sizes of transmitted frames for statistics so the latency
spike should be acceptable. In case user sets the ring sizes
to something crazy, limit the completion to 2k entries.
The check if the ring is empty at the beginning of xdp_complete()
is no longer needed - the callers will perform it.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski [Tue, 16 May 2017 00:55:20 +0000 (17:55 -0700)]
nfp: add CHECKSUM_COMPLETE support
Introduce NFP_NET_CFG_CTRL_CSUM_COMPLETE capability and implement parsing
of CHECKSUM_COMPLETE metadata.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Edwin Peer <edwin.peer@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Edwin Peer [Tue, 16 May 2017 00:55:19 +0000 (17:55 -0700)]
nfp: version independent support for chained RSS metadata
ABI version 4 introduced metadata chaining. Using the ABI version to signal
metadata chaining precludes firmware that advertises new capabilities which
rely on prepended metadata from working on older kernels.
Capability bits are thus better suited to signalling the chained metadata
format. A new version of the RSS capability is introduced to distinguish
between the differing metadata formats for ABI versions other than 4.
Signed-off-by: Edwin Peer <edwin.peer@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski [Tue, 16 May 2017 00:55:18 +0000 (17:55 -0700)]
nfp: don't assume RSS and IRQ moderation are always enabled
Even if capability for RSS and IRQ moderation are present we may
have not initialized them for control vNIC. Depend on selected
features mask (ctrl) rather than capabilities (cap) to determine
which features should be enabled.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Edwin Peer [Tue, 16 May 2017 00:55:17 +0000 (17:55 -0700)]
nfp: support LSO2 capability
Firmware advertising the LSO2 capability exploits driver provided L3 and L4
offsets in order to avoid parsing packet headers in the TX path. The vlan
field in struct nfp_net_tx_desc is repurposed, making TXVLAN a mutually
exclusive configuration to LSO2.
Signed-off-by: Edwin Peer <edwin.peer@netronome.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Edwin Peer [Tue, 16 May 2017 00:55:16 +0000 (17:55 -0700)]
nfp: rename l4_offset in struct nfp_net_tx_desc to lso_hdrlen
The l4_offset field referred to by NFD is confusingly named. It is not the
offset of the L4 transport header, but rather the L4 payload.
The LSO2 capability supported by alternative device firmware requires
the actual L4 offset, thus the rename seems prudent.
Signed-off-by: Edwin Peer <edwin.peer@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski [Tue, 16 May 2017 00:55:15 +0000 (17:55 -0700)]
nfp: don't enable TSO on the device when disabled
We advertise TSO to the stack but leave it disabled by default.
Make sure it's not only disabled in the netdev features but
also on the device itself.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
linzhang [Mon, 15 May 2017 02:26:47 +0000 (10:26 +0800)]
net: socket: mark socket protocol handler structs as const
Signed-off-by: linzhang <xiaolou4617@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Haiyang Zhang [Fri, 12 May 2017 19:14:11 +0000 (12:14 -0700)]
tools: hv: Add clean up for included files in Ubuntu net config
The clean up function is updated to cover duplicate config info in
files included by "source" key word in Ubuntu network config.
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Shannon Nelson [Wed, 10 May 2017 01:30:12 +0000 (18:30 -0700)]
bnxt: add dma mapping attributes
On the SPARC platform we need to use the DMA_ATTR_WEAK_ORDERING attribute
in our Rx path dma mapping in order to get the expected performance out
of the receive path. Adding it to the Tx path has little effect, so
that's not a part of this patch.
Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com>
Reviewed-by: Tushar Dave <tushar.n.dave@oracle.com>
Reviewed-by: Tom Saeger <tom.saeger@oracle.com>
Acked-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 16 May 2017 15:41:11 +0000 (11:41 -0400)]
Merge branch 'xgene-Add-ethtool-stats-and-bug-fixes'
Iyappan Subramanian says:
====================
drivers: net: xgene: Add ethtool stats and bug fixes
This patch set,
- adds ethtool extended statistics support
- addresses errata workarounds
- fixes bugs related to statistics
v2: Address review comments from v1
- Adds lock to protect mdio-xgene indirect MAC access
- Refactors xgene-enet indirect MAC read/write functions
- Uses mdio-xgene MAC access routines, if xgene-enet port
use the same HW.
v1:
- Initial version
Signed-off-by: Iyappan Subramanian <isubramanian@apm.com>
Signed-off-by: Quan Nguyen <qnguyen@apm.com>
====================
Signed-off-by: David S. Miller <davem@davemloft.net>